Open infrarenal abdominal aortic aneurysm repair: The Cleveland Clinic experience from 1989 to 1998

Transcription

1 CLINICAL RESEARCH STUDIES Open infrarenal abdominal aortic aneurysm repair: The Cleveland Clinic experience from 1989 to 1998 Norman R. Hertzer, MD, a Edward J. Mascha, MS, b Mathew T. Karafa, MS, b Patrick J. O Hara, MD, a Leonard P. Krajewski, MD, a and Edwin G. Beven, MD, a Cleveland, Ohio Purpose: The purpose of this study was to determine the safety and durability of traditional surgical treatment for asymptomatic infrarenal abdominal aortic aneurysms (AAAs) in a large series of patients who underwent open operations during the decade preceding the commercial availability of stent graft devices for endovascular AAA repair. Methods: From 1989 to 1998, 1135 consecutive patients (985 men [87%], 150 women; mean age, 70 7 years) underwent elective graft replacement of infrarenal AAA. Computerized perioperative data have been supplemented with a retrospective review of hospital charts/outpatient records and a telephone canvass to calculate survival rates and the incidence rate of subsequent graft-related complications. Seventy-four patients (6.5%) were lost during a median follow-up period of 57 months for the entire series. Results: The 30-day mortality rate was 1.2%. The hospital course was completely uneventful for 939 patients (83%), and the median length of stay for all patients was 8 days. A total of 196 patients had single (n 150; 13%) or multiple (n 46; 4%) postoperative complications, which were more likely to occur in men (odds ratio [OR], 2.3; 95% confidence interval [CI], 1.1 to 5.2) and in patients with a history of congestive heart failure (OR, 3.7; 95% CI, 1.7 to 7.8), chronic pulmonary disease (OR, 1.9; 95% CI, 1.2 to 2.9), or renal insufficiency (OR, 2.5; 95% CI, 1.3 to 4.7). Kaplan-Meier method survival rate estimates were 75% at 5 years and 49% at 10 years. As was the case with early complications, the long-term mortality rate primarily was influenced by age of more than 75 years (risk ratio [RR], 2.2; 95% CI, 1.7 to 2.8) or previous history of congestive heart failure (RR, 2.1; 95% CI, 1.3 to 3.4), chronic pulmonary disease (RR, 1.5; 95% CI, 1.2 to 2.0), or renal insufficiency (RR, 3.2; 95% CI, 2.2 to 4.6). Of the 1047 patients who survived their operations and remained available for follow-up study, only four (0.4%) have had late complications that were related to their aortic replacement grafts. Conclusion: These results reconfirm the exemplary success of open infrarenal AAA repair. The future of endovascular AAA repair is exceedingly bright, but until the long-term outcome of the current generation of stent grafts is adequately documented, their use should be justified by the presence of serious surgical risk factors. (J Vasc Surg 2002;35: ) Half a century has passed since DuBost, Allary, and Oeconomos 1 described the first successful replacement of an infrarenal abdominal aortic aneurysm (AAA) with a homograft. Since that time, progress in the surgical treatment of AAAs has been marked by remarkable advances in graft materials, operative techniques, and anesthesia management. Classic references, such as those by Szilagyi et al 2 and Crawford et al, 3 defined the appropriateness and the declining mortality rate of elective AAA repair, and recent data have again confirmed that spontaneous rupture is associated with such a low chance for survival that preemptive intervention is justified for AAAs of appropriate size From the Department of Vascular Surgery a and the Department of Biostatistics and Epidemiology, b The Cleveland Clinic Foundation. Competition of interest: nil. Reprint requests: Norman R. Hertzer, MD, Department of Vascular Surgery (S-61), The Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH ( Hertzen@ccf.org). Copyright 2002 by The Society for Vascular Surgery and The American Association for Vascular Surgery /2002/$ /1/ doi: /mva even in the presence of medical comorbidities. 4,5 In this regard, studies by Cronenwett et al 6,7 have suggested that hypertension and chronic obstructive pulmonary disease (COPD), rather than representing surgical contraindications, actually encourage further aneurysm enlargement and eventual rupture. The latest contribution to the treatment of AAAs endoluminal stent grafting was introduced in 1991 by Parodi, Palmaz, and Barone, 8 originally as an alternative to open repair in high-risk surgical candidates. During the decade since this landmark report was published, a growing array of proprietary stent graft devices has been used extensively in Europe and has been tested in the setting of clinical trials in the United States. Two of these (Guidant Ancure, Menlo Park, Calif; and Medtronics AneuRx, Minneapolis, Minn) have been given market approval in the United States by the Food and Drug Administration. Endovascular aneurysm repair currently is the topic of so many professional articles and symposia that vascular surgeons who do not yet perform stent grafting may have begun to wonder 1145

2 1146 Hertzer et al JOURNAL OF VASCULAR SURGERY June 2002 whether their traditional open approach already has become obsolete. Endovascular repair also has received wide attention in the lay press and on the Internet, and many patients who do not have any serious surgical risk factors seem to have the impression that stent grafting has now become the procedure of choice for all AAAs. Should endovascular AAA repair be offered routinely to patients at average or good risk? Until the late outcome of stent graft devices is better known, the answer to this question will depend in large measure on the relative risk and durability of open procedures. To document contemporary results at our own center, we have reviewed our experience with conventional AAA repair during the decade from 1989 to Table I. Patient characteristics n % Patients Men Women Age (years) Range Mean Median 70 Mean aneurysm diameter (cm) Ultrasonography (n 689) CT (n 828) Hypertension Diabetes 79 7 Insulin-dependent 15 1 Clinical cardiac risk factors None CAD CHF 38 3 Clinical pulmonary risk factors None COPD Renal risk factors Creatinine 2 mg/dl Creatinine 2-3 mg/dl 45 4 Creatinine 3 mg/dl 12 1 Prior dialysis 9 1 Replacement graft configuration Straight Aortobiiliac Aortobifemoral 40 3 Aortoiliac/femoral 22 2 MATERIAL AND METHODS During this study period, a total of 1293 consecutive patients underwent isolated open repair of AAAs that were located below the level of the renal arteries and were not associated with simultaneous renal, mesenteric, or femoropopliteal revascularization. Urgent procedures were necessary for 158 of these patients because of symptomatic intact (n 74) or ruptured (n 84) aneurysms. The remaining 1135 patients underwent elective operations for aneurysms measuring 5 cm or more in diameter and comprise our study group. Perioperative information was retrieved from their hospital charts, 39 (3.4%) of which no longer were available because they had been destroyed or misplaced during warehousing after 5 years of patient inactivity at our center. However, all postoperative deaths or complications and the essential elements of the preoperative history, the clinical cardiopulmonary status, and baseline renal function were prospectively recorded in our departmental computer registry. The only data that are missing for these 39 patients are whether they underwent functional myocardial imaging or cardiac catheterization or both before surgery. Our practice customarily is to see patients with uncomplicated conditions at 1 year and then every 3 to 5 years with a contrast-enhanced computed tomographic (CT) scan of the descending thoracic aorta and the visceral aortic segment to rule out the possibility of new proximal aneurysms. Complete follow-up information was collected on the basis of a review of the outpatient records and a yearlong telephone canvass of patients, their surviving family members, or their referring physicians. Seventy-four patients (6.5%) were lost during a median follow-up interval of 57 months (mean, months) for the entire series. Patient characteristics. As indicated in Table I, this series includes 985 men (87%) and 150 women (13%) with age ranging from 44 to 94 years (mean, years; median, 70 years). The mean diameter of AAA was cm with ultrasonography and cm with CT scanning. Eight hundred and seventeen patients (74%) had hypertensive conditions (blood pressure, 180/90 mm Hg) or needed antihypertensive medications, and only 79 (7.0%) had diabetes. Coronary artery disease (CAD) was suspected clinically in 735 patients (65%) because of either a convincing history of previous angina pectoris or myocardial infarction (MI) or the presence of a Q-wave or ischemic ST-T wave changes on a standard 12-lead electrocardiogram. Thirty-eight patients (3.3%) had a history of congestive heart failure (CHF). Clinical evidence of COPD was present in 275 patients (24%) but was investigated with functional testing only when it obviously was disabling. The preoperative serum creatinine level was 2 mg/dl or more in 57 patients (5.0%), and another nine patients (0.8%) needed maintenance hemodialysis. Preoperative cardiac assessment. We have believed for many years that optimization of the preoperative cardiac status contributes substantially to the safety of elective vascular surgery Accordingly, we routinely perform some type of objective cardiac screening before elective AAA repair unless a specific reason exists not to do so, such as previous satisfactory testing within the previous year, a relatively recent history of successful coronary artery bypass grafting (CABG) or percutaneous transluminal coronary angioplasty (PTCA), or the presence of documented CAD for which no corrective intervention is feasible. The cardiac data that are presented in Table II were obtained within 6 months before the index AAA procedure, an interval that captures all routine preoperative studies but also includes patients who may have first been discovered to have AAAs at the time they were referred to our center because of CAD. Dipyridamole-thallium scanning or dobutamine-stress echocardiography was obtained in 636 of the 1135 patients (56%). Sixteen patients underwent both of these studies to

3 JOURNAL OF VASCULAR SURGERY Volume 35, Number 6 Hertzer et al 1147 Table II. Preoperative cardiac status n % Functional myocardial imaging* Normal Fixed scar Active ischemia Inconclusive 31 5 Coronary angiography Normal 15 3 Mild to moderate CAD Advanced but compensated CAD Severe correctable CAD Severe uncorrectable CAD 4 1 Previous coronary bypass surgery Remote ( 6 months) Recent incidental Planned preliminary Previous transluminal coronary angioplasty Remote ( 6 months) Recent incidental 4 4 Planned preliminary *Dipyridamole-thallium scintigraphy (n 376) or dobutamine-stress echocardiography (n 276). Failed to attain at least 85% of predicted maximal heart rate during stress testing. clarify their findings, and 23 other patients underwent only a stress electrocardiogram. Active myocardial ischemia was identified in 104 of the 636 patients (16%) who had functional imaging studies. Coronary angiography was performed in 529 patients (47%), revealing severe but correctable CAD that appeared to place viable myocardium at risk in 154 (29%). 9 Only 121 of the 1135 patients (11%) were not known to have had either functional myocardial imaging or coronary angiography. A total of 174 patients (15%) underwent CABG (n 119; 10%) or PTCA (n 55; 4.8%) during the 6 months preceding their AAA operations, including 153 of the 154 patients who had severe correctable CAD. (The other 20 patients had miscellaneous indications for coronary intervention. Nine of them underwent PTCA [n 8] or CABG [n 1] for advanced but compensated CAD, including three who had active myocardial ischemia on functional imaging. The results of coronary angiography were unknown for six patients who underwent incidental PTCA [n 3] or CABG [n 3] elsewhere, and CABG was performed for two patients in conjunction with replacement of either the aortic valve or the ascending thoracic aorta. The remaining three charts were unavailable for review.) One hundred and eighteen of these 174 procedures (68%) were done as a direct result of the preoperative cardiac evaluation, and 56 patients had incidental AAAs that were repaired during staged procedures after primary coronary intervention. Another 332 patients (29%) had a remote history of unrelated CABG or PTCA or both. Surgical considerations. All AAA procedures were performed with general endotracheal anesthesia, usually together with epidural analgesia for postoperative pain management. Although a midline transperitoneal incision was used preferentially, an extraperitoneal approach through the left flank was used whenever it seemed to be more appropriate because of truncal obesity, serious COPD, or the possibility that suprarenal cross clamping might be necessary for construction of a juxtarenal aortic graft. (Our database does not contain specific information regarding the choice of incisions.) Collagen-impregnated polyester fiber (Dacron) replacement grafts were implanted in 1127 patients, and the remaining eight patients received polytetrafluoroethylene grafts. Straight grafts were feasible in 337 patients (30%), but aortobiiliac (n 736), aortobifemoral (n 40), or aortoiliac/femoral (n 22) bifurcation grafts were used in most cases (70%; Table I). Statistical analysis. Baseline preoperative data included gender, age, and conventional risk factors (Table I), the information obtained with noninvasive cardiac testing or coronary angiography (Table II), and a number of miscellaneous considerations, such as the year of operation, whether the AAA procedure represented a reoperation, and the configuration of the aortic replacement graft. These features were investigated for their possible relationship to postoperative outcomes, including 30-day mortality, longterm survival, and a total of 30 potential perioperative complications that are designated on our departmental registry form. These complications were subclassified in the following principal categories: cardiac, pulmonary, renal, gastrointestinal, stroke, reconstruction-related (eg, bleeding, thrombosis, infection), wound-related (eg, hematoma, infection, dehiscence), and miscellaneous (eg, amputation, deep vein thrombosis, sepsis). Perioperative complication rates were assessed primarily with logistic regression analysis. Kaplan-Meier method estimates of survival rate distribution were generated for overall survival rate and for survival rate by age, cardiac risk, prior CHF, COPD, renal function, and all other baseline factors. The relationships between these factors and survival rate were assessed with log-rank testing. Risk factors that were found to be significant were considered for multivariable analysis with the Cox proportional hazards model. Risk ratios and their 95% confidence intervals (CIs) indicate the risk at any point in time for the index group in comparison with the reference set. A significance level of.05 was used for each hypothesis. All analyses were performed with SAS statistical software (Cary, NC). RESULTS Postoperative complications. Fourteen deaths occurred within the first 30 postoperative days, for an operative mortality rate of 1.2%. Three of these deaths were cardiac-related (Table III). Three other deaths were related to pulmonary complications, four to multisystem organ failure, and the remaining four to miscellaneous causes. Nine hundred and thirty-nine patients (83%) had no complications of any kind, and the median hospital length of stay was 8 days (quartiles: 7 days, 11 days) for the entire series. The median postoperative length of stay declined from 10 days in 1989 to 8 days in 1998 (P.001), and the median length of the entire hospitalization (including preoperative days) declined from 13 days in 1989 to 8 days in

4 1148 Hertzer et al JOURNAL OF VASCULAR SURGERY June 2002 Table III. 30-day mortality and complication rates n % Deaths Cardiac Pulmonary Multisystem organ failure Other causes Any complication None Single Multiple 46 4 Cardiac complications Arrhythmia 38 3 MI 16 1 CHF 10 1 Pulmonary complications Pneumonia 35 3 Adult respiratory distress syndrome 13 1 Pulmonary embolism Renal complications Nonoliguric 11 1 Oliguric (New dialysis) (6) (0.5) Miscellaneous complications Wound 38 3 Intestinal obstruction 14 1 Intestinal ischemia 11 1 Sepsis Retroperitoneal bleeding Stroke Lower extremity DVT Amputation DVT, Deep vein thrombosis (P.001). Single complications occurred in 150 patients (13%), and 46 patients (4.1%) had multiple complications (Table III). Cardiac complications were most common (5.2%), but most of these (38/59) were confined to arrhythmias that responded to medical management. At least one set of postoperative cardiac isoenzymes was routinely obtained in most patients, and 16 perioperative MIs (1.4%) were documented. Postoperative pulmonary events occurred in 4.1% of patients, wound complications in 3.3%, and renal insufficiency (defined as an increase in the serum creatinine level of 1 mg/dl or more in comparison with the preoperative value) in 1.7%. Only six patients (0.5%) who were not undergoing preoperative dialysis needed new dialysis after surgery, however, and the incidence rate of most other serious complications was similarly low. Fortyfive early reoperations were necessary in 28 patients (2.5%), the most frequent indication being a fascial wound dehiscence (n 15; 1.3%). Significant results of the multivariable analysis of 30- day complications are presented in Table IV. The overall complication rate was higher in men and in patients who had a previous history of CHF, COPD, renal insufficiency, or dialysis. Advancing age contributed to the incidence of cardiac complications. Otherwise, the incidence of organspecific (ie, cardiac, pulmonary, or renal) complications was predictably related to preoperative risk factors, such as CHF, remote CABG, COPD, and severe renal dysfunction. Table IV. Multivariable analysis of 30-day complications Significant risk factors Odds ratio 95% CI P value All complications Men , Prior CHF , COPD , Creatinine 2 mg/dl or prior , dialysis Cardiac complications Decile age increments , Prior CHF , Remote coronary bypass surgery , Pulmonary complications COPD , Renal complications Creatinine 2 mg/dl or prior dialysis , Late survival. In addition to the 14 postoperative deaths, another 300 of the 1185 patients (25%) eventually died at a mean follow-up interval of months (median, 38 months). The principal causes of these 314 deaths were cardiac in 23% (MI, 12%; CHF, 6%; arrhythmias, 5%), cancer in 20%, pulmonary in 5%, renal in 4%, and stroke in 3%. Miscellaneous events were responsible for 4% of all deaths, but the cause of death was unavailable for 130 patients (41%). Kaplan-Meier method estimates of longterm survival rates for the entire series of patients are 94% (95% CI, 93% to 95%) at 1 year, 75% (95% CI, 72% to 78%) at 5 years, and 49% (95% CI, 43% to 54%) at 10 years (Fig 1, A). Late survival rate has been significantly influenced by age (Fig 1, B; P.001), by a clinical history of CAD (Fig 2, A; P.028), and especially by a history of previous CHF (Fig 2, B; P.001). Survival rates also have been less favorable in patients who had preoperative COPD (Fig 3, A; P.001) or renal risk factors (Fig 3, B; P.001), defined as an elevated serum creatinine level ( 2 mg/dl) or chronic renal failure that already necessitated dialysis. Additional Kaplan-Meier method survival rate estimates are summarized in Table V. The difference between the 10-year survival rates for men and women did not attain statistical significance (P.072). For reasons that are not clear, significant differences in late survival rate appear to be associated with the configuration of the aortic replacement graft (P.03) and with whether patients sustained postoperative wound complications (P.019). Overall survival rate trends among subsets of patients who underwent preoperative functional myocardial imaging (P.015) or coronary angiography (P.29) are difficult to interpret because all but one of the 154 patients who were found to have severe correctable CAD underwent CABG or PTCA within the 6 months preceding their aneurysm procedures. Thirty of the 104 patients (29%) who had active myocardial ischemia on functional imaging received some form of coronary revascularization, a feature that may have contributed to their good long-term survival rate. The survival rates for 121 patients who were not known to have had

5 JOURNAL OF VASCULAR SURGERY Volume 35, Number 6 Hertzer et al 1149 Fig 1. Kaplan-Meier method 10-year survival rate estimates for (A) all patients and (B) age quartiles. either functional imaging or coronary angiography were 90% at 1 year, 74% at 5 years, and 32% at 10 years. Significant results of the multivariable analysis of all deaths are presented in Table VI. In addition to graft configuration, the same factors that are shown with the data in Table IV to be associated with higher postoperative complication rates advanced age and previous CHF, COPD, or renal insufficiency also have been the principal predictors of late death in this series. Late graft complications. With the exclusion of the 14 patients who died after surgery and the 74 patients who eventually were lost to follow-up study, 1047 of the 1135 patients in this series were eligible for consideration regarding late complications of their aortic replacement grafts. On the basis of a review of current outpatient records and a year-long telephone canvass to obtain information for patients who had not recently undergone reexamination at our center, we were able to identify only four late graft complications (0.4%). These complications included two graft infections, one graft limb occlusion, and one femoral pseudoaneurysm in a patient for whom an aortobifemoral graft had been constructed during AAA repair. With the concession that we do not conduct routine surveillance of infrarenal aortic grafts with ultrasonography or any other imaging technique, we are not aware that any of these 1047 patients has necessitated a reoperation for a proximal aortic pseudoaneurysm. DISCUSSION Endovascular repair represents a dramatic technical advance in the management of infrarenal AAAs and already provides a relatively safe alternative to traditional open operations in truly high-risk surgical candidates. Cuypers et al 13 recently concluded, for example, that endovascular

6 1150 Hertzer et al JOURNAL OF VASCULAR SURGERY June 2002 Fig 2. Kaplan-Meier method 10-year survival rate estimates on basis of (A) any cardiac risk factor and (B) previous CHF. repair may have a lower risk for perioperative myocardial ischemia because of significant intraoperative differences in the cardiac index and the left ventricular stroke work index compared with open procedures. On the basis of their phase II clinical trials, both of the stent graft devices that have received Food and Drug Administration market approval in the United States can be deployed successfully in most patients with suitable aortic anatomy (Guidant Ancure, 94%; Medtronics AneuRx, 98%) and are associated with a low periprocedural mortality rate of approximately 2%. 14,15 Nevertheless, the durability of aortic stent grafts has not yet been determined, particularly with respect to the long-term implications of intrasac endoleaks caused by insecure attachment sites (type I), retrograde flow in branch vessels (type II), or midgraft defects (type III). Centers that participated in the phase II AneuRx trial reported early endoleaks in 38% of patients, and although only about a third of these persisted for longer than 1 month on follow-up imaging, new endoleaks eventually occurred in another 9% of patients and secondary procedures to correct endoleaks (4%) or graft limb occlusions (2%) ultimately were necessary in a total of 6% of patients. 15 According to Zarins et al, 16,17 the presence of an endoleak has not influenced the cumulative 1-year survival rate (96%) or the 2-year aneurysm rupture rate (3%) in the phase II AneuRx trial. In Europe, however, both type I and type III endoleaks have significantly contributed to the cumulative rupture rate of 1% per year for AAAs treated with several proprietary stent grafts in the EUROSTAR registry. 18 The primary outcome success rate, defined as freedom from death, aneurysm rupture, conversion to open repair, or secondary procedures for endoleak or graft occlusion,

7 JOURNAL OF VASCULAR SURGERY Volume 35, Number 6 Hertzer et al 1151 Fig 3. Kaplan-Meier method 10-year survival rate estimates on basis of (A) COPD and (B) any renal risk factor. was 88% at a follow-up interval of 18 months for 398 patients in the phase II AneuRx trial. 16 The 30-day technical success rate was only 72% for the first 1554 patients in the EUROSTAR database, but this term was stringently defined in the sense that it included deployment failure and all endoleaks in addition to death, aneurysm rupture, and secondary transabdominal intervention. 19 The EURO- STAR collaborators subsequently reported an annual open conversion rate of 2.1% during the initial 4 years of follow-up study in a larger cohort of 2464 patients, with an operative mortality rate of 24% for these 41 procedures. 18 None of these studies can anticipate whether such factors as proximal aortic dilation or material fatigue will cause additional stent graft complications in the future, but this uncertainty is another feature that calls into question whether endovascular AAA repair should presently be considered for patients who have no particular contraindications to a standard surgical approach. Precedence exists in surgical technology for concerns of this kind. Prosthetic cardiac valves, for example, have undergone many refinements in a continuing attempt to reduce their late complication rate. Perhaps the most important difference in any simile between these devices and aortic stent grafts is the fact that there was no therapeutic alternative for the replacement of diseased valves at the time that the original caged-ball prosthesis was introduced, whereas a safe, dependable operation already is available for infrarenal AAAs, especially in patients at average or good risk. How safe is open AAA repair specifically in these two groups of patients? There is no easy answer because nearly the entire literature on the topic of asymptomatic AAAs, including our report, is on the basis of nonuniform patient

8 1152 Hertzer et al JOURNAL OF VASCULAR SURGERY June 2002 Table V. Additional Kaplan-Meier survival estimates n Survival estimates 1 year 5 years 10 years P value Gender.072 Men % 75% 48% Women % 75% 58% Functional myocardial imaging*.015 Normal % 81% 63% Fixed scar % 80% 34% Active ischemia % 72% 64% Inconclusive 31 97% 88% 56% Not done % 71% 46% Coronary angiography.29 Normal or mild to moderate CAD % 72% 58% Advanced but compensated CAD % 70% 44% Severe correctable CAD % 79% 46% Not done % 79% 50% Replacement graft configuration.03 Straight % 75% 51% Aortobiiliac % 76% 47% Aortobifemoral 40 83% 63% 50% Aortoiliac/femoral % 81% 81% Wound complications.019 No % 76% 49% Yes 38 80% 52% 41% *Dipyridamole-thallium scintigraphy (n 376) or dobutamine-stress echocardiography (n 276). Preoperative information unavailable for 39 patients. Failed to attain at least 85% of predicted maximal heart rate during stress testing. Table VI. Multivariable analysis of deaths Significant risk factors Risk ratio 95% CI P value Age 75 years , Prior CHF , COPD , Creatinine 2 mg/dl or prior dialysis , Replacement graft configuration.013 Straight , Aortobiiliac 1.0 (Reference set) Aortobifemoral , Aortoiliac/femoral , populations having a wide variety of surgical risks. Representative data for elective procedures are summarized in Table VII. Collective reviews have cited several series from referral centers in which the operative mortality rate (2%) is comparable with our own (1.2%), and the mortality rate in multicentered cooperative studies generally has been about 5% Statewide audits have documented early mortality rates exceeding 7% for nonruptured AAAs but also have repeatedly shown that these rates are inversely related to the hospital volume and the experience of individual surgeons with aortic surgery The Dartmouth Atlas of Vascular Health Care has confirmed that this relationship between case volume and early outcome exists nationwide in the Medicare population, irrespective of the specialty designation of the surgeons performing the operations. 32 Recent investigations with the National Hospital Discharge Survey have the advantage of an extraordinarily large database but necessitate considerable editing even to discriminate between suprarenal and infrarenal aortic aneurysms Possibly because of differences in the way in which data were retrieved from the National Hospital Discharge Survey, these investigations have estimated that the overall mortality rate for AAA repair in the United States ranges as much as from 4% to 8%. Comparison of any of these figures with the postoperative mortality rate for endovascular AAA repair at vetted trial centers could be misleading because, in addition to potential variability in patient mix, the mortality rate of endovascular repair may prove to be much higher once this technique becomes as widely available as conventional surgical treatment. There is no immediate reason to assume, for instance, that lowvolume stent grafting will be any safer in the future than low-volume open operations have been in the past. One of the most gratifying aspects of open AAA repair is its durability. We were able to identify only four late graft complications (0.4%) among 1047 eligible patients in our

9 JOURNAL OF VASCULAR SURGERY Volume 35, Number 6 Hertzer et al 1153 Table VII. Representative operative mortality rates Year Study period Operations (n) Mortality rate (%) Collective reviews Hollier, Taylor, and Ochsner Ernst Zarins and Harris Multicenter experience Canadian Aneurysm Study Veterans Administration United Kingdom Small Aneurysm Trial Statewide audits New York Michigan Maryland National Hospital Discharge Survey Lawrence et al , Heller et al , Huber et al , series. Plate et al 33 reported a similar low incidence rate of 2.3% in 1087 referral patients at the Mayo Clinic in 1985, and, at the same center, Hallett et al 34 subsequently found that the cumulative 5-year incidence rate of major graft complications was only 7% in a population study (Olmsted County, Rochester, Minn) of patients who underwent open AAA procedures as long ago as The long-term technical results of open AAA repair historically have been so dependable, in fact, that they now are taken almost for granted. Only scant long-term data are available for endovascular repair, but Ohki et al 35 have reported a reintervention rate of 10% in 239 patients, a few of whom had been followed for as long as 9 years. Consequently, a consensus exists that aortic stent grafts must be reassessed with CT scanning at least every 6 months to detect graft migration, to evaluate old endoleaks, to discover new endoleaks, and to be certain that no interval expansion of the excluded aneurysm sac has occurred. 36 This surveillance also adds to the base cost of endografting, which, largely because of the price of stent grafts themselves, already exceeds that of open repair, despite a shorter length of stay during the index hospital admission. 37,38 Our perspective concerning the appropriateness of open AAA repair in patients at average or good risk is not prejudiced by inexperience with aortic stent grafting at the Cleveland Clinic. In the years 1999 and 2000, two new members of our department (Daniel G. Clair, MD, and Roy K. Greenberg, MD) primarily were responsible for a total of nearly 400 endograft AAA procedures during a period of time in which approximately 300 open infrarenal AAA operations also were performed. Many of the patients who underwent endograft repair during these 2 years did not have compelling contraindications to standard surgical management. Some of these patients were enrolled in ongoing device trials, but the choice of endovascular treatment in others simply reflects the fact that there is no unanimous opinion concerning the selection of patients for endograft repair even among our own staff. Despite its availability, however, the authors have confined their intradepartmental referrals for stent grafting almost exclusively to patients who have severe cardiopulmonary disease or another relative indication, such as truncal obesity, a hostile abdomen from multiple previous operations, or extreme old age. We generally have found that patients at average or good risk who specifically request endograft repair frequently have learned about this less invasive approach from the media or their personal physicians, were unaware that it had complications of its own, and often will choose to proceed with a traditional open procedure once they have been informed regarding its established safety and durability. The data in our report indicate once again that advanced age or a previous history of CHF, COPD, or renal insufficiency is associated with a significantly higher incidence rate of postoperative complications or late death after open AAA repair. Other investigators have made the same observations. 39,40 All of these findings collectively suggest to us that, until its long-term reliability has been established, aortic stent grafting is best suited to patients who are advanced in age or have truly serious medical comorbidities. The definition of what constitutes a comorbidity of sufficient severity to satisfy cautious indications for endovascular AAA repair undoubtedly will be refined as the late results of the present generation of stent grafts become clear and as future improvements in their technology make these devices even more dependable. On the basis of their initial success, it is entirely possible, perhaps even likely, that aortic stent grafting ultimately will replace open repair as the procedure of choice for infrarenal AAAs in all patients. However, this transition should be driven by longterm data proving that the benefit of endovascular procedures extends well beyond the relatively short period of time that is necessary for most patients at average or good risk to recover from a conventional operation. These data simply are not yet available. Until they are, it is unreasonable to abandon an open surgical approach that, with experience, can be performed with attainable operative mortality and late graft complication rates of 2% or less.

10 1154 Hertzer et al JOURNAL OF VASCULAR SURGERY June 2002 We thank Becky Roberts who maintained the computer database for this study. We also thank Geoffrey S. Cox, FRACS (Melbourne, Australia), and Timothy M. Sullivan, MD (Greenville, SC), who contributed to this series while they were members of our department from 1990 to 1992 and from 1994 to 1998, respectively. REFERENCES 1. DuBost C, Allary M, Oeconomos N. Resection of an aneurysm of the abdominal aorta: reestablishment of the continuity by a preserved human arterial graft, with result after five months. Arch Surg 1952;64: Szilagyi DE, Smith RF, DeRusso FJ, Elliott JP, Sherrin FW. Contribution of abdominal aortic aneurysmectomy to prolongation of life. Ann Surg 1966;164: Crawford ES, Saleh SA, Babb JW III, Glaeser DH, Vaccaro PS, Silvers A. Infrarenal abdominal aortic aneurysm. Factors influencing survival after operation performed over a 25-year period. Ann Surg 1981;193: Dardik A, Burleyson GP, Bowman H, Gordon TA, Williams GM, Webb TH, et al. Surgical repair of ruptured abdominal aortic aneurysms in the state of Maryland: factors influencing outcome among 527 recent cases. J Vasc Surg 1998;28: Noel AA, Gloviczki P, Cherry KJ Jr, Bower TC, Panneton JM, Mozes GI, et al. Ruptured abdominal aortic aneurysms: the excessive mortality rate of conventional repair. J Vasc Surg 2001;34: Cronenwett JL, Murphy TF, Zelenock GB, Whitehouse WM Jr, Lindenauer SM, Graham LM, et al. Actuarial analysis of variables associated with rupture of small abdominal aortic aneurysms. Surgery 1985;98: Cronenwett JL, Sargent SK, Wall MH, Hawkes ML, Freeman DH, Dain BJ, et al. Variables that affect the expansion rate and outcome of small abdominal aortic aneurysms. J Vasc Surg 1990;11: Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann Vasc Surg 1991;5: Hertzer NR, Beven EG, Young JR, O Hara PJ, Ruschhaupt WF III, Graor RA, et al. Coronary artery disease in peripheral vascular patients: a classification of 1000 coronary angiograms and results of surgical management. Ann Surg 1984;199: Hertzer NR, Young JR, Beven EG, O Hara PJ, Graor RA, Ruschhaupt WF, et al. Late results of coronary bypass in patients with infrarenal aortic aneurysms: the Cleveland Clinic study. Ann Surg 1987;205: Starr JE, Hertzer NR, Mascha EJ, O Hara PJ, Krajewski LP, Sullivan TM, et al. Influence of gender on cardiac risk and survival in patients with infrarenal aortic aneurysms. J Vasc Surg 1996;23: O Hara PJ, Hertzer NR, Krajewski LP, Tan M, Xiong X, Beven EG. Ten-year experience with abdominal aortic aneurysm repair in octogenarians: early results and late outcome. J Vasc Surg 1995;21: Cuypers PWM, Garien M, Buth J, Charbon J, Peels CH, Hop W, et al. Cardiac response and complications during endovascular repair of abdominal aortic aneurysms: a concurrent comparison with open surgery. J Vasc Surg 2001;33: Makaroun MS. The Ancure endografting system: an update. J Vasc Surg 2001;33:S Zarins CK, White RA, Moll FL, Crabtree T, Bloch DA, Hodgson KJ, et al. The AneuRx stent graft: four-year results and worldwide experience J Vasc Surg 2001;33:S Zarins CK, White RA, Hodgson KJ, Schwarten D, Fogarty TJ, for the AneuRx Clinical Investigators. Endoleak as a predictor of outcome after endovascular aneurysm repair: AneuRx multicenter clinical trial. J Vasc Surg 2000;32: Zarins CK, White RA, Fogarty TJ. Aneurysm rupture after endovascular repair using the AneuRx stent graft. J Vasc Surg 2000;31: Harris PL, Vallabhaneni SR, Desgranges P, Becquemin JP, van Marrewijk C, Laheij RJF, for the EUROSTAR Collaborators. Incidence and risk factors of late rupture, conversion, and death after endovascular repair of infrarenal aortic aneurysms: the EUROSTAR experience. J Vasc Surg 2000;32: Buth J, Laheij RJF, on behalf of the EUROSTAR Collaborators. Early complications and endoleaks after endovascular abdominal aortic aneurysm repair: report of a multicenter study. J Vasc Surg 2000;31: Hollier LH, Taylor LM, Ochsner J. Recommended indications for operative treatment of abdominal aortic aneurysms. J Vasc Surg 1992; 15: Ernst CB. Abdominal aortic aneurysm. N Engl J Med 1993;328: Zarins CK, Harris EJ Jr. Operative repair for aortic aneurysms: the gold standard. J Endovasc Surg 1997;4: Johnston KW. Multicenter prospective study of nonruptured abdominal aortic aneurysm. Part II. Variables predicting morbidity and mortality. J Vasc Surg 1989;9: Kazmers A, Jacobs L, Perkins A, Lindenauer SM, Bates E. Abdominal aortic aneurysm repair in Veterans Affairs medical centers. J Vasc Surg 1996;23: The UK Small Aneurysm Trial Participants. Mortality results for randomised controlled trial of early elective surgery or ultrasonographic surveillance for small abdominal aortic aneurysms. Lancet 1998;352: Hannan EL, Kilburn H Jr, O Donnell JF, Bernard HR, Shields EP, Lindsey ML, et al. A longitudinal analysis of the relationship between in-hospital mortality in New York state and the volume of abdominal aortic aneurysm surgeries performed. Health Serv Res 1992;27: Katz DJ, Stanley JC, Zelenock GB. Operative mortality rates for intact and ruptured abdominal aortic aneurysms in Michigan: an eleven-year statewide experience. J Vasc Surg 1994;19: Dardik A, Lin JW, Gordon TA, Williams M, Perler BA. Results of elective abdominal aortic aneurysm repair in the 1990s: a populationbased analysis of 2335 cases. J Vasc Surg 1999;30: Lawrence PF, Gazak C, Bhirangi L, Jones B, Bhirangi K, Oderich G, et al. The epidemiology of surgically repaired aneurysms in the United States. J Vasc Surg 1999;30: Heller JA, Weinberg A, Arons R, Krishnasastry KV, Lyon RT, Deitch JS, et al. Two decades of abdominal aortic aneurysm repair: have we made any progress? J Vasc Surg 2000;32: Huber TS, Wang JG, Derrow AE, Dame DA, Ozaki CK, Zelenock GB, et al. Experience in the United States with intact abdominal aortic aneurysm repair. J Vasc Surg 2001;33: Abdominal aortic aneurysms. In: Cronenwett JL, Birkmeyer JD, eds. The Dartmouth atlas of vascular health care. Hanover, NH: AHA Press; p Plate G, Hollier LA, O Brien P, Pairolero PC, Cherry KJ, Kazmier FJ. Recurrent aneurysms and late vascular complications following repair of abdominal aortic aneurysms. Arch Surg 1985;120: Hallett JW Jr, Marshall DM, Petterson TM, Gray DT, Bower TC, Cherry KJ Jr, et al. Graft-related complications after abdominal aortic aneurysm repair: reassurance from a 36-year population-based experience. J Vasc Surg 1997;25: Ohki T, Veith FJ, Shaw P, Lipsitz E, Suggs WD, Wain RA, et al. Increasing incidence of midterm and long-term complications after endovascular graft repair of abdominal aortic aneurysms: a note of caution based on a 9-year experience. Ann Surg 2001;234: Sapirstein W, Chandeysson P, Wentz C. The Food and Drug Administration approval of endovascular grafts for abdominal aortic aneurysm: an 18-month retrospective. J Vasc Surg 2001;34: Sternbergh WC III, Money SR. Hospital cost of endovascular versus open repair of abdominal aortic aneurysms: a multicenter study. J Vasc Surg 2000;31: Clair DG, Gray B, O Hara PJ, Ouriel K. An evaluation of the costs to health care institutions of endovascular aortic aneurysm repair. J Vasc Surg 2000;32: Johnston KW, Canadian Society for Vascular Surgery Aneurysm Study Group. Nonruptured abdominal aortic aneurysm: six-year follow-up results from the multicenter prospective Canadian aneurysm study. J Vasc Surg 1994;20: Berry AJ, Smith RB III, Weintraub WS, Chaikof EL, Dodson TF, Lumsden AB, et al. Age versus comorbidities as risk factors for complications after elective abdominal aortic reconstructive surgery. J Vasc Surg 2001;33: Submitted Sep 27, 2001; accepted Dec 20, 2001.